1. Field of the Disclosure
The disclosure relates generally to micromachined devices and, more particularly, to batch fabrication of micromachined devices with integrated polymer coating.
2. Brief Description of Related Technology
Advances in neural prostheses have been based on the development of microelectrodes directed to producing high fidelity sensation and selectively controlling the activity of neural ensembles. Silicon micromachined electrodes have been under development for many years, progressing from 2D passive probes to a variety of two-dimensional and three-dimensional active devices with integrated circuitry incorporated on-chip.
Such devices are also in wide use today for studying the nervous system. A rational map between each sensory or motor function and a corresponding region in the cerebral cortex is thought to exist. Electrical stimulation and recording techniques for the central nervous system have been investigated to study and eventually restore neurological or physiological functions, such as the sensations of sound and light. To these ends, the microelectrodes have been surgically implanted into tissue near target neurons for delivery of small currents or measurement of the extracellular action potentials from neural discharges. Early electrodes were made from insulated metal wires and glass micropipettes, and remain widely used.
Advances in neuroscience and neuroprosthetics have driven the development of more complex devices, such as multi-channel microelectrodes capable of accessing many different neurons simultaneously with good spatial resolution. Since the late 1960's, microelectronic thin-film techniques traditionally used to fabricate semiconductor devices have been utilized for this application to develop dense electrode arrays. Such microfabrication techniques have several potential technological advantages, including a high degree of reproducibility and precise control of the spatial positions of the electrode sites.
Recent work has been directed to fabricating thin-film electrodes in a silicon microelectrode array. As the traditional material used in semiconductor industry, silicon has been intensively characterized both electrically and mechanically. Selectively-diffused boron has thus been used as an etch stop to specify the thickness of the probes, or shanks, upon which the electrodes are disposed. The definition of the microelectrode arrays has thus relied upon the different wet etch rates (˜100:1) for silicon and boron-doped silicon with anisotropic silicon etchants, such as ethylene diamine pyrocatechol (EDP). Past silicon microelectrode devices have typically consisted of a silicon backend for handling and sharp penetrating shanks to insert into the neural tissue. Polysilicon or aluminum has been used for interconnects encapsulated by dielectric stacks of SiO2/Si3N4/SiO2. CMOS circuitry has been integrated into the silicon backend for control and signal processing functionality.
The thickness of the dielectric stacks has been adjusted for stress compensation to control stress-induced curvature. For intracortical applications, the probes are straight to facilitate penetration into the pia membrane. For cochlear implants, the probes are curled for easy insertion into the cochlea. After the deposition and patterning steps for the stimulating/recording electrode sites, the probes are released from the silicon wafer via the above-described anisotropic etching. This fabrication technique allows arbitrary shapes of probes and electrode sites to be patterned with dimensions controlled to better than ±1 μm. The etch stop is configured such that the released probes have a thickness, often as small as 12 μm, capable of accommodating the buckling strength for tissue penetration in neural applications.
One of the primary obstacles to long-term implantation of these devices has been the absence of a satisfactory mechanism for connecting the electrodes to the outside world, let alone one compatible with the wet etch-based fabrication technique described above. In one past case, the interconnections in these devices were implemented with flexible silicon ribbon cables. See, e.g., Hetke et al., “Silicon Ribbon Cables for Chronically Implantable Microelectrode Arrays,” IEEE Transactions on Biomedical Engineering, Vol. 41, No. 4, pp. 314-321 (1994). In this technique, a shallow boron diffusion is used to define a boron etch stop about 5 μm deep in the silicon substrate in order to fabricate the flexible silicon ribbon cable with conducting polysilicon interconnects insulated by the stacks of silicon dioxide and silicone nitride. The silicon ribbon cable extends from the microelectrodes to connect to a percutaneous plug in the skull for communication with an external control unit.
The same technique has been used for a foldable interconnecting structure for a low-profile probe. See Kim, et al., “A 64-Site Multishank CMOS Low-Profile Neural Stimulating Probe,” IEEE J. Solid-State Circuits, Vol. 31, pp. 1230-1238 (1996). In order to allow the dura membrane to be replaced over the electrode so that it remains free of the skull in chronic implant situations, the vertical rise of the electrode above the cortical surface must be less than 1 mm for the human brain. Otherwise, skull regrowth will cause the implanted device to become anchored to the skull. As a result, the silicon backend supporting the CMOS circuitry has been folded down at a right angle to the penetrating shanks so that it lays flat on the cortical surface after implantation. The silicon ribbon cables have provided a flexible interconnection between these two silicon components.
These silicon ribbon-based interconnections, however, give rise to complications during use of the probes. With a Young's Modulus of about 107 GPa, silicon has sufficient buckling strength to penetrate the pia membrane for insertion into the neural tissue during device implantation. But silicon is a rigid solid material easily fractured under stresses, and structural flexibility in connection with the silicon microelectrode may involve bending to small radii without fracture and tethering.
In addition to their susceptibility to shear stress, silicon ribbon cables are very elastic and tend to spring back to their original state after being bent. As a result, the tethering associated with bending the cable after surgical insertion can cause dislocation of penetrating shanks from the original implant area and undesirable stress on the dura membrane on top of the implanted device.
Biocompatible polymers have also been incorporated into MEMS devices, including silicon microelectrode devices. The resulting devices are generally capable of significant deformation without fracturing, and therefore can better adjust to brain micromotion with less tethering to the neural tissue.
Unfortunately, polymer materials have exhibited a tendency to develop pinholes and become embrittled after exposure to silicon etching solutions under traditional conditions for device release. As a result, integrated silicon/polymer fabrication processes have generally not been explored. Instead, alternative methods have been employed, such as post-process polymer coating, using separate polymer structures to form a hybrid connection with silicon devices, or changing the process flow to use dry etch techniques. The former two methods disadvantageously require the handling of individual structures, while the latter cannot be applied in general cases.
One technique for forming polymer-coated microelectrode devices that was not suitable for batch fabrication, or wafer-level processing, involved rivet-bonding a flexible polyimide cable to individually released silicon probes. Hetke, et al., “3-D Silicon Probe Array with Hybrid Polymer Interconnect for Chronic Cortical Recording,” Conf. Proc. First Intl. IEEE EMBS Conference on Neural Engineering, pp. 181-184 (2003). Another technique involved depositing a conformal polymer coating on the released silicon probes to form flexible structures. Drawbacks of this technique include the lack of polymer deposition selectivity and the difficulty in removing unwanted polymer coating on the electrodes. For example, laser ablation has been employed to remove the unwanted coating of iridium sites after parylene deposition. See Weiland, et al., “Recessed Electrodes Formed by Laser Ablation of Parylene Coated, Micromachined Silicon Probes,” Proc. of the 19th Annual Intl. Conf. of the IEEE Engineering in Medicine and Biology Society, Vol. 5, pp. 2273-2276 (1997). Complications arise with parylene residue leftover in under-ablated regions or iridium damage in over-ablated areas. Another effort attempted to use discharge breakdown to remove a parylene coating on the silicon probes. Akamatsu, et al., “Fabrication and Evaluation of a Silicon Probe Array on a Flexible Substrate for Neural Recording,” Proc. of the 25th Annual Intl. Conf. of the IEEE Engineering in Medicine and Biology Society, Vol. 4, pp. 3802-3805 (2003). However, the removal was undesirably limited only to the probe tips.
In each of these polymer-based techniques, post-fabrication processes required handling of individual probes. As a result, the use of polymer coatings would introduce dramatic inefficiencies once hundreds, if not thousands, of silicon probes are released from a silicon wafer.